mlsanityData to 14 Oct 2026

iMatcher: Improve matching in point cloud registration via local-to-global geometric consistency learning

arXiv 2509.08982 ↗First posted 10 Sep 2025cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

This paper presents iMatcher, a fully differentiable framework for feature matching in point cloud registration. The proposed method leverages learned features to predict a geometrically consistent confidence matrix, incorporating both local and global consistency. First, a local graph embedding module leads to an initialization of the score matrix. A subsequent repositioning step refines this matrix by considering bilateral source-to-target and target-to-source matching via nearest neighbor search in 3D space. The paired point features are then stacked together to be refined through global geometric consistency learning to predict a point-wise matching probability. Extensive experiments on real-world outdoor (KITTI, KITTI-360) and indoor (3DMatch) datasets, as well as on 6-DoF pose estimation (TUD-L) and partial-to-partial matching (MVP-RG), demonstrate that iMatcher significantly improves rigid registration performance. The method achieves state-of-the-art inlier ratios, scoring 95% - 97% on KITTI, 94% - 97% on KITTI-360, and up to 81.1% on 3DMatch, highlighting its robustness across diverse settings.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 30 of its 51 references that are in the map. First posted on arXiv 10 Sep 2025.

Builds on

30 of 51 references.

References · 30

  1. 2017Attention Is All You Needcited by 60,361
  2. 2016PointNet: Deep Learning on Point Sets for 3D Classification and Segmentationcited by 5,248
  3. 2018Dynamic Graph CNN for Learning on Point Cloudscited by 2,223
  4. 2019KPConv: Flexible and Deformable Convolution for Point Cloudscited by 1,205
  5. 2023LightGlue: Local Feature Matching at Light Speedcited by 590
  6. 2021KITTI-360: A Novel Dataset and Benchmarks for Urban Scene Understanding in 2D and 3Dcited by 575
  7. 20163DMatch: Learning Local Geometric Descriptors from RGB-D Reconstructionscited by 336
  8. 2019Deep Closest Point: Learning Representations for Point Cloud Registrationcited by 321
  9. 2020PREDATOR: Registration of 3D Point Clouds with Low Overlapcited by 290
  10. 2018BOP: Benchmark for 6D Object Pose Estimationcited by 254
  11. 2019PointNetLK: Robust & Efficient Point Cloud Registration using PointNetcited by 233
  12. 2022Geometric Transformer for Fast and Robust Point Cloud Registrationcited by 202
18 more
  1. 2020RPM-Net: Robust Point Matching using Learned Featurescited by 182
  2. 2020D3Feat: Joint Learning of Dense Detection and Description of 3D Local Featurescited by 177
  3. 2021CoFiNet: Reliable Coarse-to-fine Correspondences for Robust Point Cloud Registrationcited by 152
  4. 2021Variational Relational Point Completion Networkcited by 104
  5. 2020Fast and Robust Iterative Closest Pointcited by 67
  6. 2022SC^2-PCR: A Second Order Spatial Compatibility for Efficient and Robust Point Cloud Registrationcited by 62
  7. 2021Deep Hough Voting for Robust Global Registrationcited by 53
  8. 2023RegFormer: An Efficient Projection-Aware Transformer Network for Large-Scale Point Cloud Registrationcited by 37
  9. 2023SE(3) Diffusion Model-based Point Cloud Registration for Robust 6D Object Pose Estimationcited by 33
  10. 2024PARE-Net: Position-Aware Rotation-Equivariant Networks for Robust Point Cloud Registrationcited by 18
  11. 2022MatchNorm: Learning-based Point Cloud Registration for 6D Object Pose Estimation in the Real Worldcited by 13
  12. 2022KSS-ICP: Point Cloud Registration based on Kendall Shape Spacecited by 10
  13. 2021Robust Partial-to-Partial Point Cloud Registration in a Full Rangecited by 9
  14. 2023RDMNet: Reliable Dense Matching Based Point Cloud Registration for Autonomous Drivingcited by 5
  15. 2024Point Tree Transformer for Point Cloud Registrationcited by 5
  16. 2025Unlocking Generalization Power in LiDAR Point Cloud Registrationcited by 5
  17. 2024LoGDesc: Local geometric features aggregation for robust point cloud registrationcited by 2
  18. 2023SGNet: Salient Geometric Network for Point Cloud Registrationcited by 1

Built on by

Not cited yet.